Cycloreversion reactions of phenylated cage compounds induced by

results onf firm^^^*^ the theoretical prediction5" that oxidation of pyridine or imidazole CO complexes of iron(I1) isobacteriochlorins proceeds via t...
0 downloads 0 Views 408KB Size
670

J. Am. Chem. SOC.1981, 103, 670-672

to catalyze the six-electron reductions of nitrite and sulfite to ammonia and hydrogen sulfide, respectively. A combination of experiment and theory suggests54 that, because of the ease of oxidation of the isobacteriochlorin skeleton, a radicals play a role in the enzymatic cycles of nitrite and sulfite reductases. ESR results onf firm^^^*^ the theoretical prediction5" that oxidation of pyridine or imidazole CO complexes of iron(I1) isobacteriochlorins proceeds via the macrocycle and unambiguously establish the existence of iron a cation radicals. (54) (a) Richardson, P. F.; Chang, C. K.; Spaulding, L. D.; Fajer, J. J . Am. Chem. SOC.1979,101, 7736-8. (b) Chang, C. K.; Fajer, J. fbid. 1980, 102,848-51. (c) Richardson, P. F.; Chang, C. K.; Hanson, L. K.;Spaulding, L. D.; Fajer, J. J . Phys. Chem. 1979, 83, 3420-4.

Note Added in Proof: Hoffman and co-workerssshave obtained 'H and I4N ENDOR hyperfine splittings which further support the assignment of H R P I to a a cation. The isotopic substitution experiments suggested above for N M R are obviously equally applicable for ENDOR. Acknowledgment. We thank the authors of ref 6, 7, 9, 44,49, and 55 for providing us with their results prior to publication. This work was supported by the Division of Chemical Sciences, U.S. Department of Energy, Washington, D.C., under Contract No. DE-AC02-76CH00016 at Brookhaven National Laboratory and by N S F Grant C H E 7815285 at Michigan State University. (55) Roberts, J. E.; Hoffman, B. M.; Rutter, R.; Hager, L. P. J . Biol. Chem., in press.

Communications to the Editor Cycloreversion Reactions of Phenylated Cage Compounds Induced by Electron Transfer'

Table I. Charge-Transfer Band Maxima of 1 and 2 with TCNE in CH,CI,

x I\

Toshio Mukai,* Katsuhiro Sat& and Yoshiro Yamashitat Photochemical Research Laboratory Faculty of Science, Tohoku University Aramaki, Sendai 980, Japan Received June 30, 1980

In addition to the ring-opening reactions of cage compounds catalyzed with acid2 and metal ions,' photochemically and thermally induced valence isomerizations between cyclic dienes and cage compounds are well d o c ~ m e n t e d . ~We , ~ have found that electron transfer both in the ground and excited states also can induce cycloreversion of certain cage compounds possessing phenyl groups. In this connection there are reported examples of ring-opening reactions of monocyclic phenylated cyclobutaned and cyclopropanes7 which proceed via electron-transfer reactions. However, our findings can provide a new procedure for the cycloreversion reactions of cage compounds which occur with extraordinary high efficiency. The observed reactions are divided into three types: (1) ring opening reactions induced by electron transfer or charge transfer in the dark; (2) reactions induced by irradiation of the chargetransfer complexes; (3) reactions which occur by electron transfer between the ground state of the cage compounds and the excited state of the electron acceptors. We describe here the relationship between the types of reaction and the electron-donating or -acUniversity, Minami-Josanjima, Tokushima 770, Japan. (1) Organic Photochemistry. 47. Part 46: K. Okada and T. Mukai, Tetrahedron Lett., 359 (1980). (2) (a) T. Iwakuma, K. Hirao, and 0. Yonemitsu, J. Am. Chem. SOC., 96, 2570 (1974); (b) W. G. Dauben and L. N. Reitman, J . Org. Chem., 40,841 (1975). (3) (a) M. Sohn, J. Blum, and J. Halpern, J. Am. Chem. SOC., 101,2694 (1979); (b) P. E. Eaton and D. R. Patterson, ibid., 100, 2573 (1978); (c) P. E. Eaton and U. R. Chakraborty, ibid., 100, 3634 (1978); (d) L. A. Paquette, R. A. Boggs, W. B. Farnham, and R. S. Beckley, ibid., 97, 1112 (1975); (e) L. A. Paquette, R. A. Boggs, and J. S. Ward, ibid., 97, 1118 (1975). (4) G. Jones, 11, and B. R. Ramachandran, J . Org. Chem., 41,798 (1976). (5) T. Mukai and Y . Yamashita, Tetrahedron Lett., 357 (1978). (6) (a) P. Beresford, M. C. Lambert, and A. Ledwith, J . Chem. SOC.C, 2508 (1970); (b) A. Ledwith, Arc. Chem. Res., 5, 133 (1972); (c) T. Majima, C. Pac, A. Nakasone, and H. Sakurai, J. Chem. SOC.,Chem. Commun. 490 (1978); (d) T. Majima, C. Pac, J. Kubo, and H. Sakurai, Tetrahedron Lett., 377 (1980). (7) (a) Th. Martini and J. A. Kampmeier, Angew. Chem., fnt. Ed. Engl., 9, 236 (1970); (b) D. R. Arnold and R. W. R. Humphreys, J . Am. Chem. Soc., 101,2743 (1979); (c) S.S.Hixon, Org. Photochem., 4, 191 (1979); (d) P. C. Wong and D. R. Arnold, Tetrahedron Lett., 2101 (1979).

0002-7863/8 1/1503-670$01.00/0

2

1

W 3 R

X'

x2

1, n m

2, n m

700 (596Y 630

740 (627)'' 680 640 627 (547)'' 601

a

Me

OMe

OMe

b c

Me Me Me

OME

H

H H

OMe

Me C0,Me

c1 H

c1

d e

t Department of Applied Chemistry, Faculty of Engineering, Tokushima

I \

f

590 538

H H

" Chloranil instead of TCNE.

4 2 0 sh

endb

End absorption, 500-600 nm.

cepting properties of substrates or electron acceptors. The mechanism of type 3 is mainly described here. The charge-transfer absorption bands resulting from the interaction of bis(homocubane) derivatives la-P with tetracyanoethylene (TCNE) are shown in Table I, together with those of dienes 2a-f with TCNE. Although the chargetransfer absorption band of 4-phenylhomocubane with T C N E has been r e p ~ r t e d , ~ Table I reveals that the substitution of p-anisyl for phenyl, especially in l a and lb, results in large red shifts of the absorption maxima. This is ascribed to a through-spacei0 or through-bond (8) Synthetic method of 1 and 2 has been reported. Cf. T. Tezuka, Y. Yamashita, and T. Mukai, J . Am. Chem. SOC.,98, 6051 (1976). (9) N. A. Clinton, R. S. Brown, and T. G. Traylor, J . Am. Chem. SOC., 92, 5228 (1970). (10) R. Filler and E. W. Choe, J. Am. Chem. SOC.,91, 1862 (1969).

0 1981 American Chemical Society

J. Am. Chem. SOC.,Vol. 103, No. 3, 1981 611

Communications to the Editor

Table 11. Reduction Potential and Excited Triplet Energy of Sensitizers ET,

sensitizer anthracene acetophenone benzophenone benzil 9,lO-phenanthraquinone p-benzoquinone 9,lGanthraquinone

I

I

400

800

I

1200

[la]-‘

1600

I

2000

M-l

Figure 1. Plots of [@(Za)]-’ vs. [la]-’. Acetonitrile solutions containing 9,1&phenanthraquinone, 4.1 X 1W5M, 412 10 nm irradiation (O), and p-benzoquinone, 8.0 X lo4 M, 436 & 10 nm irradiation (O),at 23 OC.

*

interaction which is maximized in the bis-arylated cage compounds because of good overlap of the ?r orbitals of two parallel aryl groups with the u orbitals of the CI-C2 bond.” When a methylene chloride solution of a 1:l mixture of l a and TCNE was stirred at room temperature for 2 h, a 1:l adduct 3a, mp 153-155 OC, was obtained in 70% yield.I2 Under the same conditions, diene 2a added to T C N E to give 3a in 84% yield. Other p-anisyl derivatives l b and ICsimilarly react with TCNE to give dienes 2b and 2c, but only 2b affords adduct 3b, mp 167-169 OC.I3 The structural assignments of 3a and 3b are based on spectral propertiesI4 as well as the chemical evidence that heating of the adducts in refluxing benzene regenerates 2a and 2b. The fact that 2a and 2b, containing the cyclobutene ring with the anisyl group, add to TCNE also allows a structural discrimination between the two isomeric dienes Zb and 2c and between the two isomeric cage compounds l b and IC. As shown in Table I, this bis-arylated cage compound also exhibits charge-transfer absorption bands with quinones such as p-chloranil and 2,6-dichloro-p-benzoquinone.When a methylene chloride or acetonitrile solution of p-anisyl derivative la, l b , or ICand one of the above quinones was allowed to stand at room temperatures, the cycloreversion reactions giving dienes were observed without the formation of the [2 + 21 adduct. These cycloreversion reactions of la-c are classified as type 1 reactions. The diphenyl derivative Id showed no cycloreversion reactions on treatment with TCNE, p-chloranil, or 2,6-dichloro-p-benzoquinone in the dark. However, irradiation of a methylene chloride solution of Id and TCNE or p-chloranil by using a Na lamp (590 nm) for 1 h produced diene 2d in almost quantitative yields. These ring-opening reactions of Id belong to type 2 reactions. The quantum yield for the type 2 reaction of Id and TCNE ([ld] and M) was about 5 at 468 f 10 nm at room [TCNE] = 3.20 X temperature^.'^ The fact that the quantum yield is more than A. Kanematsu, et al., J. Am. Chem. SOC.,in press. (12) Satisfactory elemental analyses were also obtained for 3. (13) A,mong many examples of the [2 + 21 cycloaddition reactions between electron-rich olefins and TCNE, there is an example of hindered alkoxy styrenes which fails to add with TCNE. Cf. J. K. Williams, D. W. Wiley, and B. C. McKusick, J. Am. Chem. SOC.,84, 2210 (1962). (14) Spectral data of 3a are as follows: A, (CH,CN) 248 nm (log c 4.28); u,, (KBr) 2230 cm-I; ‘H NMR (100 MHz, CDCI,) 6 1.37-1.85 (m, 5 H), 1.82 (s, 3 H), 1.89 (s, 3 H), 2.61 (m, 1 H), 3.15 (m, 2 H), 3.57 (s, 3 H),3.72 (s, 3 H), 6.33-7.17 (m, 8 H); m / e 372 (loo%), 128 (61%). (1 5) Even if the exact quantum yield value could not be determined because of the change of the charge-transfer absorption band between cage compound and TCNE according to the progress of reaction, it is considered that this value at 5% conversion is approximately correct, since the low conversion of the cage compound causes minor change of the charge-transfer absorption band. (16) G . Briegleb, Angew. Chem., 76, 326 (1964). ( 1 7) C. K. Mann and K. K. Barnes, “Electrochemical Reactions in Nonaqueous Systems”, Marcel Dekker, New York, 1970. ( 1 1) E. Osawa,

Ered l f l r Ewed1 / 2 ,

kc4

-1.94 -1.99 -1.72 -1.04b -0.6

Va -0.10 1.22 1.28 1.31b 1.52

69.2‘ 54.3‘ 48.Sd

-0.48 -0.83

1.82 1.89

53d 62.7d

Va

mol 42.7‘ 74.1‘

reaction no Slow no

no yes yes yes

E*redl,l + sensitizer triplet energy; sensitizer groundstate reduction potentials (vs. SCE, DMF) are complied in ref 17. Value in Me,SO complied in ref 17. ‘Values complied in S. L. Murov, “Handbook of Photochemistry”, Marcel Dekker, New York, 1973. Values complied in S . Patai, “The Chemistry of Functional Groups. The chemistry of the quinonoid compounds”, Interscience, New York, 1974.

unity suggests that a chain reaction should be involved in type 2 reactions. With weak electron acceptors such as p-benzoquinone, 9,lOanthraquinone, 9,1O-phenanthraquinone, s-trinitrobenzene, and p-dicyanobenzene, neither cage compounds l a 4 nor dienes 2a-d exhibit any charge-transfer absorption band. However, when an acetonitrile solution of l a or Id containing p-benzoquinone or 9,lO-phenanthraquinone was irradiated with light of longer wavelength than 400 nm for a short time (1-1 5 min), diene 2a or 2d was obtained in good yields (70-90%). These reactions belong to type 3, and occur very efficiently with increasing quantum yields as the concentration of l a or Id is increased. Surprisingly the limiting quantum yields of the reactions were found to be infinite, as shown in Figure 1, suggesting the occurrence of a chain reaction. The reactions were quenched with piperylene. In addition, the luminescence of 9,lO-anthraquinone1* was quenched by l a with the rate k, = 1.4 X lo8 M-’ s-I. In order to establish the limitation of type 3 reactions, the photosensitized reactions of l a were carried out in the presence of other sensitizers such as benzophenone, acetophenone, benzil, and anthracene. Although a slow conversion to 2a took place when acetophenone was used,I9 other sensitizers did not lead to the reaction. As shown in Table 11, the reactions depend on the reduction potentials but not the triplet energies of the sensitizers, except for acetophenone. Thus, a classical energy-transfer mechanism is less likely than an electron-transfer mechanism. In connection with this, the oxidation potentials of cage and diene compounds were polarographically measured and the EoXlI2values of 1.09 and 1.41 V (vs. SCE) for l a and Id and 1.09 and 1.40 V for 2a and 2d were obtained.20 Comparison of these potential values with the reduction potentials (ErCdlj2in Table 11) of the electron acceptors’’ along with the fact that the type 3 reaction of l a was quenched with triethylamine (Eox112 = 0.82 V vs. SCE) should support the electron-transfer mechanism. On the basis of these findings, we propose a reaction mechanism for the type 3 reactions which involves a chain process. The

(18) S . A. Carlson and D. M. Hercules, J . Am. Chem. SOC.,93, 5611 (1971). (19) In this case, the exothermic energy-transfer reaction may occur. (20) Half-wave potentials vs. SCE in acetonitrileusing 0.1 M Et4NtC104as supporting electrolyte; 0.1 V s-’ scan speed. We thank Professor Nobuyuki Tanaka and Dr. Akifumi Yamada for providing the measurements.

672

J. Am. Chem. SOC.1981, 103, 672-673

initiation step consists of excitation of the quinone and generation of the radical cation 1+. by electron transfer from the cage compounds. The ring cleavage process of the radical cation 1+. giving the radical cation 2+., followed by a subsequent electron transfer from 1 to 2+., comprises the propagation steps. Then, the radical cation 1’. reenters the chain cycle. Reaction 4 can be considered to be reversible, because there is no gap in the oxidation potentials between 1 and 2. However, reaction 3 involving the bond cleavage reaction releases much strain energy and cannot be reversible. Thus, the propagation sequence is promoted. The termination step involves a quenching reaction of the radical cations 1’. and 2’. with the radical anion Q-..The absence of sensitizer radical ion from the propagation sequence probably allows the long chain lengths observed. The detailed mechanism for the type 1 and 2 reactions is still apbiguous, but the key step must be electron transfer from the cage compounds to the electron acceptors in both cases. Since the type 2 reaction also involves a chain process, there seems no essential differentiation between the type 2 and 3 reactions. When the electron-transfer reaction occurs thermally from the cage compounds to T C N E or quinone, type 1 reaction might result. On the other hand, if the reaction requires excitation of charge-transfer complexes or sensitizers, type 2 and 3 reactions will occur. More recently, we have discovered that the type 3 reactions of Id also take place with quantum yields over unity when dye sensitizers such as triphenylpyrylium and trityl salts2’ or semiconductors such as ZnO and CdS were used.22 Thus, it could be concluded that electron-releasing properties of the cage compounds which are ascribed to their strain energies and oxidation potentials might be important factors for the cycloreversion reactions which are induced to occur by electron transfer.

Acknowledgment. Financial support was provided from the Ministry of Education, Science and Culture. (21) K. Okada, K. Hisamitsu, T. Miyashi, and T. Mukai, to be published. (22) K. Okada, K. Hisamitsu, and T. Mukai, J. Chem. Soc., Chem. Commun., in press.

Statistical State Solvation Sites P. K. Mehrotra, Francis T. Marchese, and David L. Beveridge* Department of Chemistry Hunter College of the City Univrsity of New York New York,New York 10021 Received July 31, 1980 The structural organization of solvent molecules around a dissolved solute in solution has been discussed extensively in the recent literature in terms of solvation sites. In aqueous solutions, solvation sites are the regions in which water molecules are likely to be found in the vicinity of a dissolved molecular solute. The solvation site concept can be extended straightforwardly to molecular solutions in general. Up to this point, solvation sites have been equivalenced with local minima in the solutesolvent potential energy hypersurface, as advanced by Pullman, Pullman, and co-workers.’ Calculations (1) A. Pullman and B. Pullman, Q. Rev. Biophys., 7 , 505 (1975).

0002-7863/81/1503-672$01.00/0

of solvation sites based on molecular quantum mechanics have been reported for diverse biological molecules and used to discuss aqueous hydration and solvent effects on biomolecular conformational structure. The Pullman approach has generated useful insight into aqueous solvation processes but can be improved upon by including solventsolvent interactions, temperature effects, and consideration of the statistical weight of solvation sites in the N-particle molecular assembly.2 We offer here a new definition of solvation sites expressed on the statistical state of an N-molecule system at a given temperature in which all of the above critical points are accommodated. The procedure is illustrated by the determination of solvation sites for trans-glyoxal from an analysis of computer simulation results on the dilute aqueous solution, [(HCO)2]aq,at 25 “C. All information about the structure of molecular solutions is in principle contained in the generic molecular distribution functions for the system, and information on solvation sites follows in principle from the analysis of these functions in the region of the first solvation shell of a solute. A general procedure for the analysis of the local solution environment of a dissolved solute from molecular distribution functions has recently been proposed from this l a b o r a t ~ r y . ~Here the solvent molecules in the various N-molecule configurations of the system are assigned to the closest or most proximal solute atom (the proximity criterion). The subsequent analysis is organized in the general framework of quasicomponent distribution function t h e ~ r y . ~By use of this approach the primary and higher order solvation of each solute atom is determined. The spatial extent of the first coordination shell of the solute can be well defied from the solutesolvent radial distribution for primary solvation. The probability density distribution for solvent molecules in the first shell can thus be determined, and the spatial probability distribution of solvent molecules in the immediate vicinity of the solute can be displayed. The further resolution of this density according to solute atoms or functional groups can be achieved by using the proximity criterion, and the analysis of these results leads to the determination of solvation sites. These sites are defined on the statistical state of the system with N-molecule interactions, temperature factors, and statistical weights taken into account and are henceforth called “statistical state solvation sites”. Alternative uses of probability density maps to analyze simulation results on aqueous solutions have been described by Hagler et a1.5 and Romano and ClementL6 To examine the viability of the statistical state solvation site concept, we have carried out a liquid state ( T , V, N) ensemble Monte Carlo computer simulation on [(HC0)2]aq,represented by 1 glyoxal molecule and 21 5 water molecules under periodic boundary conditions at 25 OC. Ensemble averages in the calculation of properties and in the structural analyses were formed from lOOOK configurations after equilibration, chosen on the basis of the Metropolis method. General aspects of the calculations and related recent results are reviewed in ref 7 and 8 and literature cited therein. The primary coordination number of glyoxal in [(CH0)2]aq was found to be 17.77 water molecules. The 90% probability density distribution of these molecules was determined and displayed by using computer graphics by means of the program PSI/77 by Jorgensen? The results are collected in the composite Figure 1.

(2) D. L. Beveridge and G. W. Schnuelle, J. Phys. Chem., 78,2064 (1974). (3) P. K.Mehrotra and D. L. Beveridge, J . Am. Chem. Soc., 102, 4287 (1980). (4) A. Ben-Naim, “Water and Aqueous Solutions”, Plenum Press, New York, 1974. ( 5 ) A. T. Hagler, D. J. Osguthorpe, and B. Robson, Science 298, 599 (1980). ( 6 ) S. Romano and E. Clementi, Inr. J. Quantum Chem., 17, 1007 (1980). (7) D. L. Beveridge, M. Mezei, S. Swaminathan, and S. W. Harrison, ACS Sym. Ser., No.86, 191 (1978). (8) D. L. Beveridge, M. Mezei, P. K. Mehrotra, F. T. Marchese, V. Thirumalai, and G. Ravi-Shanker, Ann. N.Y.Acad. Sci., in press. (9) W. L. Jorgensen, QCPE, 11, 340 (1979).

0 1981 American Chemical Society